Description
Zorro is an AI testing platform for teams shipping software quickly and wanting confidence that new code hasn't broken existing features. The product, previously called Narrative, is Y Combinator-backed and offers two connected capabilities: AI bug bash and AI regression testing.
Read More
AI bug bash uses agents that explore an application the way a manual QA tester would, proactively clicking through flows to surface unexpected edge cases, content mismatches, broken layouts, visual flaws, and functional bugs. Because the agents figure out paths dynamically rather than following a fixed script, there are no test cases to write or update as the UI changes. Every bug found comes with a full recording, console logs, and a step-by-step breadcrumb trail so engineering and product teams can reproduce and fix it without guesswork.
AI regression testing lets teams write tests in plain English instead of maintaining XPath or CSS selectors. Zorro's AI assertions evaluate against the intent of a test rather than a fixed value, which the product positions as better suited to complex product logic and dynamic content. Runtime autohealing resolves selector changes, network flakiness, and cookie banners automatically, so a UI tweak like a renamed button doesn't break a whole suite; when autohealing can't confidently recover, it's flagged for review rather than silently passing. Tests can be created directly from a coding agent through Zorro MCP, and both bug bash and regression testing can run before a pull request merges or on a recurring schedule.
Zorro runs entirely in its own cloud infrastructure, so there are no npm packages to install, browser extensions to add, or local runners to configure, and a planning agent can spin up multiple parallel browsers to cut down exploratory testing time. It supports web, iOS, and Android. On security, tests run in isolated containers, Zorro states it does not train public AI models on customer DOM, screenshots, or test data, secrets are encrypted at rest with AES-256, and a SOC 2 audit is in progress.
Pricing starts at $1,000 per month, with shared Slack support included for all customers and enterprise plans adding SSO and priority support; exact tiers require contacting Zorro directly.
- CategoryAI Code Review & Testing
- Pricing ModelPaid
- DeploymentCloud
- PlatformsWeb, iOS, Android
Zorro's Core Features
AI bug bash agents that explore an app like a manual QA tester
AI regression tests written in plain English instead of code selectors
AI assertions that evaluate against test intent rather than fixed values
Runtime autohealing for selector changes, flakiness, and cookie banners
Full recordings, console logs, and breadcrumbs attached to every bug found
Parallel agent execution across multiple browsers to speed up testing
Zorro MCP for creating tests directly from a coding agent
Cloud-hosted execution with no local runners or browser extensions to manage
Support for web, iOS, and Android testing
How to use Zorro?
Connect your app: Point Zorro at your web, iOS, or Android app and the environment to test, such as staging.
Choose a test type: Write a regression test in plain English, or launch an AI bug bash to explore proactively.
Run the agents: Zorro executes the test or exploration in its cloud, using parallel browsers where possible.
Review results: Check recordings, console logs, and breadcrumbs for any bugs or failures found.
Fix and re-run: Share findings with engineering, then re-run tests, which autoheal as the UI changes.
Automate the workflow: Trigger tests before a pull request merges or on a recurring schedule.
Zorro's Use Cases
- Pre-release regression testing
- Exploratory bug hunting
- PR-triggered testing
- Release sanity checks
